What's Changed

• Breaking change: Remove numpy np.random.RandomState usage and deprecate random_state parameter by @relf in #763
• Bump pypa/cibuildwheel from 3.1.3 to 3.1.4 by @dependabot[bot] in #751
• Bump actions/setup-python from 5 to 6 by @dependabot[bot] in #753
• Prepare release 2.10 by @relf in #764

Full Changelog: v2.9.5...v2.10.0

What's Changed

• Add a scikit-learn framework adapter for surrogate models by @Paul-Saves in #740
• Support egobox 0.32 by @relf in #752
• Make expand_lhs() function reproducible by @relf in #755
• Update adaptative DOE notebook by @mlalande

Full Changelog: v2.9.4...v2.9.5

What's Changed

• Update hierarchical variables handling documentation by @Paul-Saves in #741
• Upgrade GPX to egobox 0.31 by @relf in #749
• Bump pypa/cibuildwheel from 3.0.0 to 3.1.3 by @dependabot[bot] in #745
• Bump actions/download-artifact from 4 to 5 by @dependabot[bot] in #746
• Bump actions/checkout from 4 to 5 by @dependabot[bot] in #747
• Support Python 3.13 by @relf in #750

Full Changelog: v2.9.3...v2.9.4

What's Changed

• Add the activation vector as an input of hierarchical based functions by @Paul-Saves in #726
• Add new multi-fi cokriging with Sparse gaussian process (aka SMFCK) by @mcastanoUQ in #727
• Update the documentation and apply ruff version by @Paul-Saves in #729
• Bump pypa/cibuildwheel from 2.23.1 to 3.0.0 by @dependabot in #724, #730, #728

Full Changelog: v2.9.2...v2.9.3

What's Changed

• Add Sparse MFK application by @mcastanoUQ in #718
• Make SGP reproducible by @relf in #721
• Bump pypa/cibuildwheel from 2.22.0 to 2.23.0 by @dependabot in #716
• Bump pypa/cibuildwheel from 2.23.0 to 2.23.1 by @dependabot in #719

Full Changelog: v2.9.1...v2.9.2

What's Changed

• Fixes 3D plots with matplotlib 3.9.x by @relf in #705
• Handle scipy 1.15 changes related to derivative function used in benchmarking problems by @relf in #707
• Add pre-commit hook for ruff linting and formatting by @relf in #708
• Fix pva and add a seed for test reproducibility by @Paul-Saves in #709
• Fix kriging-based surrogates predict_variances when noise0 option is given as a np.ndarray by @relf in #710
• Allow to disable hyperparameter optimization in kriging-based surrogates by @relf in #712
• Documentation and updated notebook for MFCK by @mcastanoUQ in #714
• Prepare release 2.9.1 by @relf and @NatOnera in #713

Full Changelog: v2.9.0...v2.9.1

What's Changed

• Add save/load methods for surrogate models by @Antoine-Averland, @relf in #689, #691, #693, #694, #703
• Add the option to specify a seed for the Random sampling method by @enricostragiotti in #696
• Add metric computation functions for surrogate performance evaluation by @Antoine-Averland in #698
• Add an example of surrogate-based optimization using constrained EI by @heinehav, @Paul-Saves in #699, #702
• Add multi-fidelity adaptative doe examples by @mlalande, @NatOnera in #701
• Update the documentation on how to build the documentation by @enricostragiotti in #697
• Prepare release 2.9 by @relf in #704

New Contributors

• @enricostragiotti made their first contribution in #697
• @heinehav made their first contribution in #699
• @mlalande made their first contribution in #701

Full Changelog: v2.8.1...v2.9.0

✨ Merry ChriSMTas! ✨

What's Changed

• Fix optimized LHS (ESE) by @neal-ks in #683
• Update mixed-integer notebook to work with 2.8 by @Paul-Saves in #677
• Add ruff format checking in Github action by @Paul-Saves in #685
• Remove erroneous import from applications by @Paul-Saves in #678
• Remove zeroing small variance values by @Paul-Saves in #684
• Add a warning on kriging-based surrogates being trained with multiple outputs by @relf in #686
• Bump pypa/cibuildwheel from 2.21.3 to 2.22.0 by @dependabot in #680

New Contributors

• @neal-ks made their first contribution in #683

Full Changelog: v2.8.0...v2.8.1

What's Changed

• Add combination of kernels handling by @NicolasJeanGonel in #653
• Add implementation of the multi-fi cokriging model, MFCK (see tutorial) by @mcastanoUQ in #657
• Add save/load methods to GPX surrogate model by @relf in #658
• Move advanced design space handling in separate smt-design-space-ext module by @Paul-Saves in #651
• Remove smt dependency to ConfigSpace and adsg-core by @Paul-Saves in #663
• Refactor smt.design_space by @relf in #665, #666, #667
• Make error message to increase nugget more actionable. by @NAThompson in #661
• Fix continuous relaxation compatibility with hierarchical modeling by @Paul-Saves in #660, #664
• Update SMT tutorials, doc, affiliation by @Paul-Saves in #668, #669, #671
• Update kernel engineering notebooks by @jomorlier in #674
• Bump pypa/cibuildwheel from 2.21.1 to 2.21.3 by @dependabot in #662
• Prepare release 2.8 by @relf in #675

Implementation Note

In this release, advanced design space management has been moved in a separate smt-design-space-ext module to decouple smt from special design space implementations allowing hierarchical variables handling.

New Contributors

• @mcastanoUQ made their first contribution in #657

Full Changelog: v2.7.0...v2.8.0

What's Changed

• Add cooperative components kriging model as application (see doc and tutorial) by @lisapretsch in #642
• Refactor kernel functions in class-based kernels for kriging by @NicolasJeanGonel in #622
• Add EGO with noisy data by @MaelTremouille and @Paul-Saves in #646
• Add methods (random or kmeans) to induce points in SGP by @MaelTremouille and @relf in #633
• Add documentation for local bases by @Shisohka in #632
• Add tutorial and notebooks ruff linting by @NatOnera and @Paul-Saves in #639, #643
• Update POD+I documentation by @sdubreui and @Paul-Saves in #654
• Add test for predict_variance_derivatives in gpx by @Paul-Saves in #640
• Clean utils after switching from genn to jenn by @Paul-Saves in #626
• Fix bugs in likelihood computation by @Paul-Saves in #634
• Fix SGP for non-normalized data by @Paul-Saves in #638
• Fix continuous relaxation and hierarchical handling by @Paul-Saves in #647
• Fix is_ri computational time by @Paul-Saves in #652
• Bump pypa/cibuildwheel from 2.19.2 to 2.21.1 by @dependabot in #635, #644 and #655
• Drop Python 3.8 support by @relf in #656

New Contributors

• @lisapretsch made their first contribution in #642
• @sdubreui made their first contribution in #654

Full Changelog: v2.6.3...v2.7.0